Quality data are accurate depictions of the real world that are consistent across an enterprise, secure and accessible, delivered in a timely manner, and suitable for their intended applications (Redman 2001).
The most obvious measure of data quality is accuracy, or the degree to which the information “accurately” depicts the real world construct or phenomenon it represents. For example, a student’s numeric grade for algebra I in the LDS must match the one printed on his or her report card. To be accurate, data reported and maintained must also be complete; for instance, every student must be identified as either male or female.
Quality data must be consistent across the enterprise. For instance, a student’s name should be recorded in the same manner in every silo system. While a particular child may answer to "Charles," "Charlie," or "Chuck," only one form of his first name should be maintained by the agency. In addition, calculated data items such as he dropout rate should be computed the same way if they are calculated more than once. Internal consistency may also be referred to as "integrity," which may be compromised when data are somehow corrupted during a data transfer or other process. The concept of consistency is also similar to the that of "reliability," which may be diminished if, for instance, the definition of a data element is unclear, leaving room for varying interpretations by the staff creating the data. Additionally, data must be coded in a manner that adheres to defined code sets. Ultimately, inconsistent data will not be comparable for analysis; nor will they be easily interoperable or portable.
Quality data must be delivered within a useful timeframe. While a data system may be able to provide teachers with accurate student test scores, the data will be of limited use if they take months to deliver. Thus, while data may be considered of high quality by other measures, they must be accessible to users quickly if they are to meet their intended purpose of providing actionable information for decisionmaking.
Quality data must be secured to protect privacy and to prevent tampering (see chapter 8). At the same time, these data must be available to authorized users to provide information and improve decisionmaking.